2.3.0

What's Changed

New features

Turn on/off classification augmentations (#4039)
Turn on/off detection and instance segmentation augmentations (#4066)
Add GPU memory monitor hook (#4118)
Add YOLOv9 model for Object Detection (#3917, #4026)
Add OV inference for keypoint detection (#3970)
Add tiling for semantic segmentation (#3954)
Add D-Fine Detection Algorithm (#4142)

Enhancements

Update visual prompting pipeline for multi-label zero-shot learning support (#3993)
Update to work torch compile in detection (#4003)
Refactor MaskDINO (#4006)
Fix MaskRCNN/RTMDet-Inst/MaskRCNNTV Explain Mode (#4053)
Reduce tiling unittests resource consumption (#4073)
Upgrade OpenVINO to 2024.5 and NNCF to 2.14.0 (#4123)
Upgrade OV, MAPI, and NNCF dependencies (#3967)
Instance Segmentation Model refactoring (#3865)
Bump torch and lightning to 2.4.0 versions (#3843)
Add mAP metric to evaluate multilabel classification (#3985)
Bump OV to 2024.6, update empty label handling (#4162)
Bump MAPI to 0.2.5.1 (#4211)
Revert FMetric computation (PR: 4130) (#4210)

Bug fixes

Fix MaskDINO recipe (#4080)
Fix a wrong HPO log (#3972)
Update model name in rotated detection recipes (#4028)
Fix SupCon flag (#4076)
Add h-cls label info normalization (#4173)
Fix arrow support for semantic segmentation task (#4172)
Fix backward compatibility issues in model checkpoint loading (#4199)
Fix label info on loading checkpoint (#4200)
Revert h-cls head to linear one (#4221)

Full Changelog: 2.2.2...2.3.0

2.2.2

What's Changed

• Support Ellipse Shape for InstSeg algo by @eugene123tw in #4152
• BC improvement by @sovrasov in #4154

Full Changelog: 2.2.1...2.2.2

2.2.1

What's Changed

• Allow empty tile annotation by @eugene123tw in #4124
• Fix tensor type compatibility in dynamic soft label assigner and RTMDet head by @eugene123tw in #4140
• Update Label Info handling by @sovrasov in #4127
• Fix early stopping in converter patching + fix lr warmup for all tasks by @kprokofi in #4131
• Decouple DinoV2 for semantic segmentation by @kprokofi in #4136
• Ensure target class indices are of type long in loss calculations by @eugene123tw in #4143
• Fix arrow format reader for multiclass ROI case by @sovrasov in #4145
• Update classification in converter by @sovrasov in #4146
• Prepare 2.2.1 release by @kprokofi in #4147

Full Changelog: 2.2.0...2.2.1

2.2.0

What's Changed

• [hotfix] Update for fix workflow issues by @yunchu in #3673
• Disable integration test in pr-merge workflow by @yunchu in #3674
• Align integration test to exportable code interface update by @eunwoosh in #3675
• Merge Back Hot-Fix into develop by @harimkang in #3682
• Bump aquasecurity/trivy-action from 0.22.0 to 0.23.0 by @dependabot in #3660
• Bump actions/upload-artifact from 4.3.2 to 4.3.3 by @dependabot in #3661
• Bump openpyxl from 3.1.2 to 3.1.4 by @dependabot in #3665
• Fix segmentation fault on VPM PTQ to develop by @sungchul2 in #3690
• Mergeback 2.1.0rc0 to develop by @yunchu in #3694
• Add Multi-Label & H-label Classification with torchvision models by @harimkang in #3697
• Remove DataModuleConfig for the reduced config structure by @wonjuleee in #3688
• Fix torchvision CLS recipes by @wonjuleee in #3705
• Use model_api action classification model wrapper class by @eunwoosh in #3610
• Improve polygon utils by @wonjuleee in #3707
• Refactor Visual Prompting OV model to use MAPI by @sovrasov in #3638
• Fix multi batch testing for visual prompting by @sungchul2 in #3695
• Fix load_reference_info for zero-shot learning by @sungchul2 in #3709
• Fix a wrong action classification transform_fn for PTQ by @eunwoosh in #3713
• Mergeback 2.1.0rc0 to develop 2nd by @yunchu in #3718
• Add HuggingFace Model Wrapper for Multi-Class Classification by @harimkang in #3710
• Move intg tests to merge-check wf and run on a10g by @yunchu in #3711
• Recipe refinement (det, iseg, r-det, sseg, vpm, zsl) by @sungchul2 in #3712
• Reimplementation of ViT for classification by @wonjuleee in #3719
• Remove merge_group event by @yunchu in #3731
• Fix Combined Dataloader & unlabeled warmup loss to work correctly in Semi-SL case by @harimkang in #3723
• Update semantic segmentation transforms to use OTX's instead of torchvision by @sungchul2 in #3724
• Mergeback 2.1.0 to develop by @yunchu in #3730
• Update vpm data pipeline by @sungchul2 in #3725
• Update trivy scanning job by @yunchu in #3733
• Fix detection integration test by @sungchul2 in #3734
• Disable mixed preicision to NMS on CPU by @eunwoosh in #3727
• Enable integration test at pre-merge workflow by @yunchu in #3732
• Add groups setting to create single PR for each group by @yunchu in #3735
• Add LoRA option for VisionTransformer by @wonjuleee in #3729
• Remove optimizer in stfpm recipe by @eunwoosh in #3743
• Change run-on target to aws ci for e2e by @yunchu in #3745
• Add '--test-only' flag in perf test by @eunwoosh in #3746
• Add HuggingFace Model Wrapper for Detection by @harimkang in #3747
• Add HuggingFace Model Wrapper for Semantic-Segmentation by @harimkang in #3749
• Support dinov2 models (small, base, large, giant) from VisionTransformer backbone by @wonjuleee in #3744
• Update daily.yaml by @yunchu in #3751
• Enable to override data configurations by @sungchul2 in #3748
• Change warning to DEBUG_MODE in mem_cache by @harimkang in #3757
• Enable to use input_size at transforms in recipe by @eunwoosh in #3759
• Update recipes using $(input_size) by @sungchul2 in #3763
• Update tutorials/advanced/configuration.rst by @sungchul2 in #3765
• Add RT-DETR for Object Detection by @kprokofi in #3741
• Enable torch.complile to work with classification by @harimkang in #3758
• Update Documentation by @harimkang in #3768
• Bump actions/upload-artifact from 4.3.3 to 4.3.4 by @dependabot in #3717
• Bump ossf/scorecard-action from 2.3.3 to 2.4.0 by @dependabot in #3772
• Remove typ for torch upper version by @harimkang in #3775
• Enable to use polygon and bitmap mask as reference mask for zero-shot learning by @sungchul2 in #3769
• Add task overriding for ConfigConverter by @harimkang in #3774
• Bump ipython from 8.23.0 to 8.26.0 by @dependabot in #3701
• Bump github/codeql-action from 3.25.10 to 3.25.15 by @dependabot in #3773
• Refine recipes for anomaly tasks by @harimkang in #3777
• Refine recipes for classification by @harimkang in #3776
• Add num_devices in Engine for multi-gpu training by @harimkang in #3778
• Bump actions/setup-python from 5.1.0 to 5.1.1 by @dependabot in #3739
• Bump mlflow from 2.14.0 to 2.15.0 by @dependabot in #3771
• Upgrade NNCF and OV versions by @sovrasov in #3646
• Improve yolox-tiny tile recipe by @eugene123tw in #3767
• Introduce OTXTrainType by @harimkang in #3779
• Add otx micro benchmark by @sovrasov in #3762
• Remove mm compatible modules by @sungchul2 in #3780
• Update 'otx benchmark' print outputs & cli print outputs by @harimkang in #3782
• Enable direct export of ViTs by @sovrasov in #3736
• Bump torch version to 2.2.2 by @yunchu in #3786
• Update dependabot config by @yunchu in #3785
• Mergeback 2.1.0 to develop by @yunchu in #3787
• Refactoring base module (ConvModule) by @sungchul2 in #3783
• Change deterministic value to False as default in action classification perf test by @eunwoosh in #3799
• Update dependabot group config for pip by @yunchu in #3798
• Add keypoint detection task with RTMPose by @wonjuleee in #3781
• Bump lightning version to 2.3.3 by @yunchu in #3796
• Bump psutil from 5.9.8 to 6.0.0 by @dependabot in #3807
• Bump the github-actions-dependency group across 1 directory with 3 updates by @dependabot in #3804
• Hot-Fix hang issue in integration test by @harimkang in https://github.com/op...

2.2.0rc14

What's Changed

• Add Keypoint Detection legacy template by @kprokofi in #4094
• Revert the old workaround for detection confidence threshold by @sovrasov in #4096
• OTX RC 2.2 version up by @sovrasov in #4099
• Add dummy XAI to RTDETR (export mode) & disable strong aug by @eugene123tw in #4106
• Fix task chain for Det -> Cls / Seg by @kprokofi in #4105

Full Changelog: 2.2.0rc13...2.2.0rc14

Release 2.2.0rc13

What's Changed

• Fix wrong model name in converter & template by @harimkang in #4082
• RTMDet Inst Seg Explain Mode for 2.2 by @eugene123tw in #4083
• Fix rtdetr recipes by @sungchul2 in #4079
• Enable adaptive_bs with Efficientnet-V2-L model template by @harimkang in #4085

Full Changelog: 2.2.0rc12...2.2.0rc13

Release 2.2.0rc12

What's Changed

• Bump onnx to 1.17.0 to omit CVE-2024-5187 by @yunchu in #4063
• Fix incorrect all_groups order configuration in HLabelInfo by @harimkang in #4067

Full Changelog: 2.2.0rc11...2.2.0rc12

Release 2.2.0rc11

What's Changed

• Update anomaly transforms by @ashwinvaidya17 in #4059

Full Changelog: 2.2.0rc10...2.2.0rc11

Release 2.2.0rc10

What's Changed

• Add legacy template LiteHRNet_18 template by @kprokofi in #4049
• Model templates: rename model_status value 'DISCONTINUED' to 'OBSOLETE' by @leoll2 in #4051
• Enable export of feature vectors for semantic segmentation task by @kprokofi in #4055
• Upgrade MAPI in 2.2 by @sovrasov in #4052
• Update MRCNN model export to include feature vector and saliency map by @harimkang in #4056
• Fix applying model's hparams when loading model from checkpoint by @sungchul2 in #4057

Full Changelog: 2.2.0rc9...2.2.0rc10

Release 2.2.0rc9

What's Changed

• Fix wrong indices setting in HLabelInfo by @harimkang in #4044

Full Changelog: 2.2.0rc8...2.2.0rc9